The Internet is a set of networks connected by routers. A router maintains a routing table that indicates for each possible destination network, the next hop to which a received data packet should be forwarded. The next hop may be another router or the final destination.
An Internet Protocol (“IP”) data packet received at a port in a router includes an IP destination address. The IP destination address is the final destination of the IP data packet. Currently there are two versions of IP, IP version 4 (“IPv4”) and IP version 6 (“IPv6”). IPv4 provides a 32-bit field in an IP header included in the data packeted for storing the IP destination address. The router forwards a received data packet connected to a next-loop router, or the final destination if the destination is the local network, dependent on the IP destination address stored in the IP header.
A 32-bit IPv4 destination address provides 4 billion possible routes. An Internet router typically stores 50,000 of the 4 billion possible routes. However, the number of stored routes will increase with the growth of the Internet and the widespread use of IPv6.
Originally, the IP address space was divided into three classes of IP addresses; A, B and C. Each IP address space was divided into a network address and a host address. Class A allowed for 126 networks and 16 million hosts per network. Class B allowed for 16382 networks with 64,000 hosts per network and class C allowed for 2 million networks with 256 hosts per network. However, dividing the IP address space into different classes reduced the number of available IP addresses. Class C only allowed a maximum of 256 hosts per network which is too small for most organizations. Therefore, most organizations were assigned a Class B address, taking up 64,000 host addresses which could not be used by other organizations even if they were not used by the organization to which they were assigned. Hosts in an organization with a Class B IP address all store the same network address in the 16 Most Significant Bits (“MBSs”), for example, 128.32.xx.xx.
Classless InterDomain Routing (“CIDR”) was introduced to free up unused IP host addresses. The remaining unused networks are allocated to organization in variable sized blocks. An organization requiring 500 addresses gets 500 continuous addresses. For example, an organization can be assigned 500 available addresses starting at 128.32.xx. The number of routes stored by a router has increased since the introduction of Classless InterDomain Routing. Classless InterDomain Routing requires longest prefix matching to find the corresponding route instead of searching for a matching network address in order to find the corresponding next hop for the IP destination address. For example, a search can no longer stop after the 16 MSBs of a Class B IP address, for example, 128.xx.xx because 128.32.4.xx may be assigned to another organization requiring a different next hop.
One method for searching for a longest prefix match for a key is through the use of a binary tree search. A binary tree search matches a 32-bit input bit by bit down to 32 levels, requiring 32 searches to find the entry matching the 32-bit key. Another method for searching for a match is through the use of a Patricia tree. A Patricia tree reduces the number of searches required if there are no entries down a leaf of the binary tree.
Yet another method for efficiently searching for a next hop associated with an IP destination address is described in PCT application Serial Number PCT/SE98/00854 entitled “Method and System for Fast Routing Lookups” by Brodnick et aL filed on May 11, 1998. The method described by Brodnick reduces the number of next hops stored by not storing duplicate routes. By reducing the number of next hops, the memory requirement is reduced so that a route lookup table can be stored in fast cache memory.
Brodnick et aL divides the binary tree into 3-levels. Dividing the binary tree into 3-levels reduces the number of searches to three. The indexed entry in the first level indicates whether the search can end at the first level with the route taken from the entry, or the search must continue to a subsequent level using a further portion of the IP destination address.
A bit is set in the bit map 30 if the node does not share the route entry the previous node 44. Bit 4 is set ‘1’ indicating that there is a subtree or route index corresponding to the node 44. Bit 5 is set ‘0’ indicating that the node corresponding to bit 5 shares the subtree or route index with the node corresponding to bit 4. Bit 6 is set ‘1’ indicating that the node corresponding to bit 6 does not share a subtree or route index with bits 4 and 5. Bit 6 corresponds to a node sharing the default route. Bits 6 and 12 set to ‘1’ correspond to the default route.
The array of code words 36, array of base indices 34 and map table 40 are used to select a pointer in an array of pointers (not shown). The pointer stores a route index or an index to perform a further search.
A group of pointers in the array of pointers is selected by selecting a code word 46 in the array of code words 36 and a base index 42 in the array of base indices 34. The code word 46 is selected using the first 12 bits 50 of the IP address 38. The base index 42 is selected using the first 10 bits 48 of the IP address 38. The correct pointer in the group of pointers is selected using the map table 32.
The 10-bit value 46b in the selected code word 36 is an index into the map table 32. The map table 32 maps bit numbers within a bit-mask to 4-bit offsets. The offset specifies the pointer within the selected group of pointers in the array of pointers. The 10-bit value 46b selects the row in the map table 32 and bits 19:16 of the IP address 52 selects the 4-bit offset 54.
An offset to specify a pointer is stored in the map table 32 for each ‘1’ stored in the bit mask 30 (
Sixteen entries are used to store the eight pointers to routes r1-r8 and the default route r0. The pointer to the default route r0 is stored in eight of the sixteen entries 2521-16.
A route corresponding to a node in the subtree is found by counting the number of ‘1’s in the bit map and incrementing the pointer by the total number of ones. For example, the entry corresponding to node 224 (
A longest prefix match lookup table defining nodes of a tree searched for a route pointer corresponding to a prefix match is presented. The lookup table stores a binary tree representation of a key in a plurality of subtree levels. A portion of the bits of the key are searched in each subtree level.
Each subtree level includes a subtree memory and a mapper memory. The result of a search of each subtree level indicates whether a search must continue in a subtree in the next subtree level. The mapper memory stores pointers for nodes in a subtree. The subtree memory stores a subtree descriptor indexed by a subtree select from the previous subtree level. Instead of including a bit per node in the bottom level of the subtree, the subtree descriptor includes a bit for each node in the subtree. The bit corresponding to the node indicates whether a pointer for the node is stored in the mapper memory. By increasing the number of bits in the subtree descriptor, a pointer to a default route for the subtree shared by a plurality of nodes in the subtree is stored in a single entry in the mapper memory. Thus, the number of available locations for storing routes in the mapper memory is increased.
If the bit corresponding to a node in the subtree descriptor is set to ‘1’, a pointer is stored for the node in mapper memory. The subtree descriptor includes a level descriptor for each level in the subtree. Each level descriptor includes a bit for each node in the level. Each subtree level also includes mapper address logic. The mapper address logic performs a parallel search in each level descriptor for a node matching a search key and computes an offset to the pointer corresponding to the matching node.
The subtree descriptor may also include a block pointer. The block pointer may store an index to the first pointer in the mapper memory for the subtree. The combination of the offset and the block pointer provides the index to the pointer corresponding to the matching node.
The default route is stored for a root node of the subtree. The default route stored may be an indication to use another default route for a parent subtree.
The foregoing and other objects, features and advantages of the invention will be apparent from the following more particular description of preferred embodiments of the invention, as illustrated in the accompanying drawings in which like reference characters refer to the same parts throughout the different views. The drawings are not necessarily to scale, emphasis instead being placed upon illustrating the principles of the invention.
A description of preferred embodiments of the invention follows.
U.S. patent application Ser. No. 09/733,627 filed on Dec. 8, 2000 describes a method and apparatus for storing a route for an Internet Protocol (“IP”) address in a multi-level lookup table, the contents of which are incorporated herein by reference. A multi-level search is performed to find a route index stored in a mapper in the lookup table which indexes a range of IP addresses corresponding to a range of leaves of a subtree.
Thus, the bit map includes a bit for each node in the subtree instead of only leaf nodes as in the prior art coding scheme described in conjunction with
A bit map for a subtree of h=3 has 15 (24+1−1) instead of 8 bits 23 for a bit map which has a bit per leaf node.
The subtree data 308 includes a bit for each node in the subtree. A bit for each node in the bottom level of the N-level subtree; that is, the leaf nodes is provided in level N nodes bitmap 312. A bit for each node in the level above the bottom level (N-1) is provided in level N-1 nodes bitmap 314 and a bit for the root of the subtree is provided in the level 1 node bitmap 316.
The subtree pointers field 310 includes a default index 318 for storing the default route for the subtree and a block pointer 320 for storing a pointer to the start of a block of mapper entries allocated for storing routes for the subtree. The default index 318 is described in co-pending U.S. application Ser. No. 10/004,280 filed on Oct. 31, 2001 entitled “Default Route Coding,” by David A. Brown, the contents of which are incorporated herein by reference.
As shown in
As shown, a pointer to each route r1-r8 in the subtree shown in
Returning to
In one embodiment, the default route is stored in the default index 318 to facilitate updating the default route and allow the default route to be inherited from a parent subtree. An inherit indicator is stored in the default index 318 instead of the default route, to indicate that the default route corresponding to the root of the subtree is inherited from a parent subtree. A use default indicator is stored in mapper entry 4069 instead of the default route itself. Upon detecting the use default indicator stored in the mapper entry, the pointer to the default route stored in the default index or the inherited default route pointer is returned as the result of the search of the lookup table.
The node index 510 for a subtree of height N has N bits, for example, the node index 510 for a subtree of height 3 has 3 bits. The three bits identify the leaf node (i.e., the node at the bottom of the subtree) and all parent nodes up to the root of the subtree.
Returning to
The subtree descriptor 300 includes subtree data 308 (
The subtree data 308 stored in the subtree descriptor 300 is forwarded to the mapper address logic 504. The mapper address logic 504 also receives a node index 510. The mapper address logic 504 determines the mapper index 516 to the mapper entry 412 corresponding to the node in the subtree dependent on the node index 510, the subtree data 308 and the subtree pointers 310 in the subtree descriptor 300 for the subtree. The mapper index 516 selects the mapper entry 412 for the node in the subtree.
The total number of mapper entries lower than the mapper entry for the selected node is computed by summing the offsets 6021-602n output by the offset count logic 6001-600n for each level. The mapper index 516 is computed in the adder 610 by adding the pointer offsets to the block pointer 320 (
The level N offset logic 600n includes a level N thermometer decoder 700, a bitmap comparator 702 and a level N offset adder 704. If there is a mapper entry stored for the level N node, the level N thermometer decoder 700 selects all nodes lower than the leaf node index 604. A match N signal 708 is set to ‘0’. The match N signal is coupled to the bitmap comparator 702 in the level N-1 offset count logic 600n−1. The match N signal 708 set to ‘0’ disables the output of offset count logic 600 of lower levels. If there is no mapper entry stored for the level N node, all of the bits in the match bit map 706 are forced high. The match N signal 708 is set to ‘1’ to enable the output of offset count logic 600 of lower levels.
The bitmap comparator 702 compares each bit in the match bit map 706 with the respective bit in the level N nodes bitmap 312 stored in the subtree descriptor 300. The result of the comparison is output as the offset bitmap 710. The offset bitmap 710 indicates all mapper entries 406 that are stored in mapper memory 402 up to the selected node for the subtree. The level N offset adder 704 counts the number of ‘1’s in the offset bit map 710. The total number of ‘1’s is the level N offset 602n of the mapper entry for the node from the block pointer 320.
The level 7 thermometer decoder 8027 corresponds to the level N thermometer decoder for level N, where N=7 which was described in conjunction with
Level 7 of the subtree has 128 (27) nodes. The thermometer decoder 8027 includes a 1 of 128 decoder 808, a plurality of AND-gates, one for each of the 128 outputs of the 1 of 128 decoder 808, a NOR gate 812 and a plurality of OR gates, one for each output of the plurality of AND-gates. The bit map comparator 8047 includes a plurality of AND gates, one for each bit in the level 7 nodes bit map. The operation of the mapper address logic 504 is explained using a numerical example. The bit map (for the most significant 21 bits of the 128 bitmap) stored in subtree data 308 (
The node index [6:0] is ‘001000’ indicating a search for the route corresponding to the 17th node in the seventh level of the subtree. The level 7 nodes bit map SM70-SM7127 for the first 21 nodes in the seventh level of the subtree is shown in Table 1. The bit map SM70-SM7127 is stored in the level N (N=7) nodes bit map 312 (
Each output from the 1 of 128 decoder 808 is compared with a respective bit of the level 7 nodes bitmap SM70-SM7127 to determine if there is a mapper entry for the 17th node; that is, if there is a mapper entry for the 17th node in the 7th level of the subtree. As shown in Table 1, only bit 16 of the bit map output from the plurality of the AND gates 8100-810127 is set to ‘1’ indicating that there is a mapper entry for the 17th node. Each of the AND gate outputs is coupled to a respective input of NOR gate 812. If any of the inputs to NOR gate 812 is set to ‘1’, the output of NOR gate 812 is set to ‘0’, indicating that there is a mapper entry corresponding to a node in the 7th level. If the output of NOR gate 812 is set to ‘0’, a further search for a match in the level 6 nodes or nodes in any other upper levels of the subtree is not necessary because a node in the 7th level is selected according to the longest match requirement when there are multiple matches in the subtree.
As shown, there is a mapper entry for the 17th node in level 7 of the subtree, thus the output of the NOR gate 812 is set to ‘0’. The output of NOR gate 812 (match N 708 (
The 2-input OR gates select all bits in the level 7 nodes bitmap SM70-SM7127 that are lower than or equal to the selected node (17th node) selected by the 1 of 128 decoder 808. The match bit map 706 (
The bitmap comparator 702 includes a plurality of 2-input AND-gates 8160-816127, one for each bit in the match bit map 706. Each AND gate 8160-816127 compares one bit of the match bit map 706 with a respective bit of the level 7 nodes bitmap SM07-SM7127 312 (
The speed of the mapper address logic 504 can be improved using other components such as look ahead adders, parallel versus serial logic and adder pre-decoders to reduce propagation delay. These techniques are well-known to those skilled in the art.
At step 900, each decoder 808, 810 in each respective offset count logic (
At step 902, the node map output by the decoder is compared with the node bit map for the respective level. If any of the resulting bits are ‘1’, there is a mapper entry for the node and processing continues with step 904. If not, processing continues with step 910 to compute the number of mapper entries used by the level.
At step 904, all mapper entries lower than the selected node in the level are selected in parallel in each level offset count logic 802 by selecting all bits above the selected node and comparing with the node bit map for the respective level. Processing continues with step 906.
At step 906, each level adder 806 computes the offset based on the number of mapper entries up to the selected node in the level. Processing continues with step 908.
At step 908, adder 610 adds the total number of mapper entries from all levels stored in mapper memory for nodes up to the selected node to the block pointer to provide the mapper index 516 to the pointer for the selected node. Processing is complete.
At step 910, there is no mapper entry for the selected node. All the mapper entries for the level are added to provide the offset to the first mapper entry in the next level. Processing continues with step 908.
A lookup table including a subtree descriptor encoding a subtree using one bit per node requires less overall memory than the prior art subtree descriptor encoding a subtree using one bit per leaf node. The reduction in memory for an embodiment capable of storing 256K routes in mapper memory is shown below in Table 2.
The number of bits per entry in subtree memory is increased from 256 bits to 294 bits. The 294 bit entry includes 256 bits of subtree data, a 20-bit default index and an 18-bit block pointer field. As shown in Table 2, the total memory is reduced from 20 Mega bits to 15 Mega bits.
The invention has been described for an embodiment in which the mapper address logic is implemented using decoders, adders and combinational logic (AND, NOR, OR gates). However, in an alternate embodiment, the mapper address logic can be implemented using a Content Addressable Memory (CAM) as is well known to those skilled in the art.
While this invention has been particularly shown and described with references to preferred embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the scope of the invention encompassed by the appended claims.
This Patent Application is a Continuation of U.S. patent application Ser. No. 11/969,521 filed on Jan. 4, 2008, entitled, “DENSE MODE CODING SCHEME,” the contents and teachings of which are hereby incorporated by reference in their entirety. U.S. patent application Ser. No. 11/969,521 is a continuation of U.S. patent application Ser. No. 10/262,643 (U.S. Pat. No. 7,346,009) filed Sep. 30, 2002, entitled “DENSE MODE CODING SCHEME.”
Number | Date | Country | |
---|---|---|---|
Parent | 11969521 | Jan 2008 | US |
Child | 12612443 | US | |
Parent | 10262643 | Sep 2002 | US |
Child | 11969521 | US |